Terminal replacement prediction methods, devices, electronic equipment and storage media
By determining multi-dimensional replacement factors through operator-side data, calculating the replacement index and time window, and generating replacement prediction results, the problem of high misjudgment rate and low hit rate in existing technologies is solved, and more accurate user replacement prediction is achieved.
Patent Information
- Application Number
- CN202611484411.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-22
- Publication Date
- 2026-11-17
AI Technical Summary
Existing technologies that predict user replacement needs by setting a fixed threshold based on the age of the terminal have a high false positive rate and a low hit rate, and cannot accurately predict user replacement behavior.
By identifying multi-dimensional replacement factors, including multi-dimensional lock-in factors, using operator-side data, calculating the replacement index and replacement time window, and generating replacement prediction results, the multi-dimensional lock-in factors reflect user resistance to replacement. By combining the accuracy and coverage of operator-side data, the false positive rate is reduced and the hit rate is improved.
It enables more accurate prediction of users' device replacement time window and willingness to replace devices, reduces the misjudgment rate of prediction results, improves the hit rate, meets compliance constraints, and avoids sensitive issues.
Smart Images

Figure CN123123877A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of communication technology, and in particular to a terminal replacement prediction method, apparatus, electronic device, and storage medium. Background Technology
[0002] With the development of software and hardware technologies, mobile phones and other terminals are being updated and replaced at an increasingly rapid pace. In pursuit of better performance and a smoother experience, users' replacement needs have shifted from simply "replacing when it breaks" to prioritizing brand, performance, trendiness, and appearance. The key is to accurately predict users' replacement needs in order to better provide replacement services and improve user satisfaction.
[0003] In related technologies, a fixed threshold is set based on the age of the terminal (e.g., a terminal older than 24 months is considered a high-potential replacement user) to predict users' replacement needs.
[0004] However, the method of predicting users' replacement needs by setting a fixed threshold based on the age of the terminal has a high misjudgment rate and low hit rate. Summary of the Invention
[0005] The purpose of this application is to provide at least one terminal replacement prediction method, device, electronic device and storage medium, which can at least solve the problems of high misjudgment rate and low hit rate of prediction results, and at least achieve the effect of reducing the low misjudgment rate and improving the hit rate of prediction results.
[0006] To address the aforementioned technical problems, at least one embodiment of this application provides a terminal replacement prediction method, comprising: determining a multi-dimensional replacement factor based on operator-side data corresponding to the user's current terminal, wherein the multi-dimensional replacement factor includes a multi-dimensional locking factor reflecting the user's resistance to replacement; calculating a replacement index for the user to replace the current terminal based on the multi-dimensional replacement factor; calculating a replacement time window for the user to replace the current terminal based on the multi-dimensional locking factor, wherein the window size of the replacement time window is negatively correlated with the value of any locking factor; and generating a replacement prediction result based on the replacement index and the replacement time window.
[0007] By utilizing multidimensional replacement factors that represent replacement intentions, including multidimensional locking factors that represent replacement resistance, the final replacement prediction result is obtained. This approach characterizes user replacement behavior from two main perspectives: whether users want to replace their devices and whether they are able to replace them. Compared to using a fixed threshold based on device age to predict replacements, this method reduces the false positive rate and increases the accuracy rate of prediction results.
[0008] Furthermore, by using multidimensional locking factors to calculate the user's device replacement time window and limiting the window size of the replacement time window to be negatively correlated with the value of any locking factor, it may seem to violate the rule that the greater the user's resistance to device replacement, the larger the replacement time window. However, this is actually an accurate model of user device replacement behavior based on a large amount of data analysis, summarization, and discovery. It can more accurately represent the relationship between the user's device replacement window and the multidimensional locking factors, thereby more accurately predicting the user's device replacement time window. Furthermore, limiting the data source to operator-side data ensures comprehensive coverage, high accuracy, and compliance with regulations, thus avoiding sensitivity issues when obtaining data from other sources.
[0009] In some optional embodiments, the multidimensional locking factors include a convergence binding factor, a contract lock-in period factor, and a family ripple factor. The convergence binding factor reflects the depth of the SIM card binding between the current terminal application and the operator's converged service. The contract lock-in period factor reflects the remaining time of the operator contract bound to the current terminal. The family ripple factor reflects the strength of the mutual influence of phone-changing behaviors within the family corresponding to the current terminal. The step of calculating the phone-changing time window for a user to change the current terminal based on the multidimensional locking factors includes: calculating the phone-changing time window according to the following formula:
[0010]
[0011] in, , , The factors are, in order: fusion binding factor, contract lock-in period factor, and family ripple factor; , , In order , , The preset weight parameters; To lock in the efficiency score; The remaining time; This is the preset replacement delay baseline.
[0012] The above formula is used to calculate the device replacement time window, which provides a more specific relationship between the window size and the multidimensional locking factor. It also more specifically and accurately reflects the relationship between the multidimensional locking factor and the device replacement time window. The device replacement time window calculated based on the above formula has a higher hit rate and a lower false positive rate, which makes it easier to provide users with better device replacement services based on the device replacement time window and improve user satisfaction.
[0013] In some optional embodiments, the SIM card of the current terminal application is the primary card in the family; generating the phone replacement prediction result based on the phone replacement index and the phone replacement time window includes: converting the phone replacement index into a cluster level according to a preset first mapping rule; obtaining the family binding degree corresponding to the family and the phone replacement time window corresponding to each secondary card in the family; and calculating the ripple propagation probability of the user's phone replacement on the family members corresponding to each secondary card according to the following formula:
[0014]
[0015]
[0016] in, For the user to replace the device, the first in the household The ripple propagation probability corresponding to each secondary card; This is the ripple propagation amplification factor; The family binding degree; It is a cluster adjustment factor obtained by mapping the cluster level according to a preset second mapping rule; For the first The corresponding device replacement time window for each secondary SIM card. This refers to the device switching time window corresponding to the current terminal. For the time window difference of equipment replacement, It is the attenuation factor; The attenuation coefficient is used; the device replacement prediction result is generated based on the ripple propagation probability of the family members corresponding to each sub-card, the grouping level, and the device replacement time window of the current terminal.
[0017] For primary card scenarios, the ripple propagation probability of primary cardholder switching to secondary cardholders is introduced when generating switching prediction results. This identifies opportunities for bulk switching within a family, allowing for better switching services from both the individual user's and their family's perspectives. This improves user experience and increases user satisfaction. The formula above calculates the ripple propagation probability by multiplying the ripple propagation amplification coefficient, family binding degree, cluster adjustment factor, and attenuation factor. This method more accurately reflects the impact of primary cardholders on secondary cardholders' switching, resulting in a higher hit rate and a lower false positive rate.
[0018] In some optional embodiments, the multi-dimensional replacement factor further includes a device age factor, an average monthly user revenue factor, a data usage factor, and a terminal tier factor. The device age factor reflects the current terminal's age, the average monthly user revenue factor reflects the operator's average monthly revenue from the user, the data usage factor reflects the data usage of the current terminal, and the terminal tier factor reflects the market value of the current terminal. Generating the replacement prediction result based on the ripple propagation probability of family members corresponding to each secondary SIM card, the grouping level, and the current terminal's replacement time window includes: determining the user's replacement direction based on the device age factor, average monthly user revenue factor, data usage factor, and terminal tier factor, where the replacement direction is one of upgrading, remaining at the same tier, or downgrading; and generating the replacement prediction result based on the ripple propagation probability of family members corresponding to each secondary SIM card, the grouping level, the current terminal's replacement time window, and the replacement direction.
[0019] The clearly defined replacement factors also include device age, average monthly user income, data usage, and terminal tier, which respectively characterize the newness of the terminal, purchasing power, performance requirements, and starting tier for replacement, fully covering the core driving factors for replacement direction decisions. The replacement direction determined using these four dimensions has a high accuracy rate. Incorporating the replacement direction into the replacement prediction results helps to formulate replacement strategies that better meet user needs, thereby providing better replacement services and improving user satisfaction.
[0020] In some optional embodiments, the multi-dimensional replacement factors further include device age factor, average monthly user income factor, traffic factor, terminal tier factor, and age factor; the multi-dimensional locking factors include integration binding factor, contract lock-in period factor, and family ripple factor; and the age factor is used to reflect the user's age. The step of calculating the replacement index for the user to replace the current terminal based on the multi-dimensional replacement factors includes: performing a weighted summation of the replacement factors other than the multi-dimensional locking factors, and subtracting the weighted summation of the multi-dimensional locking factors from the resulting value to obtain a preliminary replacement index; calibrating and normalizing the preliminary replacement index using the calibration coefficient corresponding to the current terminal to obtain the final replacement index.
[0021] By using five intention-related factors (device age factor, average monthly user income factor, data traffic factor, terminal level factor, and age factor) and three lock-in factors, totaling eight dimensions, we can better reflect users' willingness and resistance to upgrading their devices. This makes the calculated device upgrade index more consistent with users' true intentions and improves the accuracy of user device upgrade predictions.
[0022] In some optional embodiments, the calibration coefficient is calculated as follows:
[0023] in, The calibration coefficient is mentioned above; This represents the cumulative replacement rate of current-generation products within the current terminal's product series. The percentage of users who have replaced their devices with the previous generation of the series to which the current terminal belongs; , All are preset values greater than 0, and ; Indicates will The value is limited to .
[0024] After weighted summation of multi-dimensional replacement factors, calibration is performed using calibration coefficients corresponding to the terminal series to eliminate differences in replacement rhythms among different brands, generations, and price ranges of terminals. This makes the calculated replacement index more in line with users' true intentions and improves the hit rate of user replacement prediction.
[0025] At least one embodiment of this application also provides a terminal replacement prediction device, comprising: a determining module, configured to determine a multi-dimensional replacement factor based on operator-side data corresponding to the user's current terminal, wherein the multi-dimensional replacement factor includes a multi-dimensional locking factor for reflecting the user's resistance to replacement; a calculating module, configured to calculate a replacement index for the user to replace the current terminal based on the multi-dimensional replacement factor; and calculate a replacement time window for the user to replace the current terminal based on the multi-dimensional locking factor, wherein the window size of the replacement time window is negatively correlated with the value of any locking factor; and a generating module, configured to generate a replacement prediction result based on the replacement index and the replacement time window.
[0026] In some optional embodiments, the multidimensional locking factor includes a convergence binding degree factor, a contract lock-up period factor, and a family ripple factor. The convergence binding degree factor reflects the depth of the SIM card binding between the current terminal application and the operator's converged service. The contract lock-up period factor reflects the remaining time of the operator contract bound to the current terminal. The family ripple factor reflects the strength of the mutual influence of handset switching behaviors within the family corresponding to the current terminal. The calculation module is used to calculate the handset switching time window according to the following formula:
[0027]
[0028] in, , , The factors are, in order: fusion binding factor, contract lock-in period factor, and family ripple factor; , , In order , , The preset weight parameters; To lock in the efficiency score; The remaining time; This is the preset replacement delay baseline.
[0029] In some optional embodiments, the SIM card of the current terminal application is the primary card in the family; the generation module is used to convert the phone switching index into a grouping level according to a preset first mapping rule; obtain the family binding degree corresponding to the family and the phone switching time window corresponding to each secondary card in the family; and calculate the ripple propagation probability of the user's phone switching on each family member corresponding to the secondary card according to the following formula:
[0030]
[0031]
[0032] in, For the user to replace the device, the first in the household The ripple propagation probability corresponding to each secondary card; This is the ripple propagation amplification factor; The family binding degree; It is a cluster adjustment factor obtained by mapping the cluster level according to a preset second mapping rule; For the first The corresponding device replacement time window for each secondary SIM card. This refers to the device switching time window corresponding to the current terminal. This refers to the time window difference for equipment replacement. The attenuation coefficient is used; the device replacement prediction result is generated based on the ripple propagation probability of the family members corresponding to each sub-card, the grouping level, and the device replacement time window of the current terminal.
[0033] In some optional embodiments, the multi-dimensional replacement factor further includes a device age factor, an average monthly user revenue factor, a data usage factor, and a terminal tier factor. The device age factor reflects the current terminal's age, the average monthly user revenue factor reflects the operator's average monthly revenue from the user, the data usage factor reflects the data usage of the current terminal, and the terminal tier factor reflects the market value of the current terminal. The generation module determines the user's replacement direction based on the device age factor, the average monthly user revenue factor, the data usage factor, and the terminal tier factor. The replacement direction is one of upgrading, remaining at the same tier, or downgrading. The replacement prediction result is generated based on the ripple propagation probability of family members corresponding to each secondary SIM card, the grouping level, the current terminal's replacement time window, and the replacement direction.
[0034] In some optional embodiments, the multi-dimensional device replacement factors further include device age factor, average monthly user income factor, traffic factor, terminal tier factor, and age factor; the multi-dimensional locking factors include integration binding factor, contract lock-up period factor, and family ripple factor; and the age factor is used to reflect the user's age. The calculation module is used to perform a weighted summation of the device replacement factors other than the multi-dimensional locking factors, and subtract the weighted summation of the multi-dimensional locking factors from the obtained value to obtain a preliminary device replacement index. The preliminary device replacement index is calibrated and normalized using the calibration coefficient corresponding to the current terminal to obtain the final device replacement index.
[0035] At least one embodiment of this application also provides an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the terminal switching prediction method described above.
[0036] At least one embodiment of this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the terminal replacement prediction method described above.
[0037] At least one embodiment of this application also provides a computer program product, including a computer program that, when executed by a processor, implements the terminal replacement prediction method described above. Attached Figure Description
[0038] One or more embodiments are illustrated by way of example with reference to the accompanying drawings, and these illustrative descriptions do not constitute a limitation on the embodiments.
[0039] Figure 1 This is a flowchart of a terminal replacement prediction method provided in one embodiment of this application; Figure 2 This is a flowchart of generating replacement prediction results provided in one embodiment of this application; Figure 3 This is another flowchart for generating replacement prediction results provided in one embodiment of this application; Figure 4 This is a schematic diagram of a terminal replacement prediction device provided in another embodiment of this application. Detailed Implementation
[0040] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided to make this application more comprehensive and complete, and to fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
[0041] Furthermore, the accompanying drawings are merely illustrative of this application and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.
[0042] Example 1: The specific implementation methods of the embodiments of this application will be described in detail below with reference to the accompanying drawings.
[0043] The terminal replacement prediction method of this embodiment can be applied to electronic devices with communication, computing, and data storage capabilities. Its specific process can be as follows: Figure 1 As shown, it includes: S101-S104.
[0044] S101, determine the multi-dimensional replacement factor based on the operator-side data corresponding to the user's current terminal. The multi-dimensional replacement factor includes a multi-dimensional locking factor used to reflect the user's resistance to replacement.
[0045] Here, "current terminal" refers to the terminal that the user is currently using, such as a smartphone or wearable device. This embodiment uses a smartphone as an example for illustration.
[0046] Operator-side data includes user and terminal-related data accumulated by telecom operators during their daily business operations. In some examples, operator-side data includes, but is not limited to: terminal billing data (e.g., recording terminal model, age, billing status, etc.), TAC (Type Allocation Code) terminal identification database (identifying terminal brand, model, and tier through the TAC code in IMEI (International Mobile Equipment Identity), contract management data (e.g., recording the type of contract a user has signed up for, contract expiration date, etc.), bundled package data (recording whether a user has signed up for broadband, TV, and mobile phone bundled services and their binding status), family association data (e.g., recording the relationship between primary and secondary SIM cards, family account member composition, member's phone replacement time, etc.), billing data (e.g., recording average monthly revenue per user (ARPU), average monthly data usage, etc.), and user profile data (e.g., recording user age, etc.). All of the above data is directly accessible from the operator's internal systems and, compared to third-party data, is characterized by comprehensive coverage, high accuracy, and strong uniqueness.
[0047] Multidimensional switching factors include a set of factors composed of multiple quantifiable factors, used to comprehensively characterize the user's switching-related features.
[0048] In some examples, the switching factor for each dimension is normalized to the corresponding numerical range (e.g., 0 to 1) according to the corresponding original data and a preset mapping rule. The numerical value is used to characterize the strength of the user's switching tendency (or switching resistance) under that dimension.
[0049] The multidimensional lock-in factor is a subset of the multidimensional replacement factor, specifically designed to reflect the level of resistance users face when replacing their devices. A higher lock-in factor value indicates greater resistance to device replacement and greater difficulty in freely replacing devices in the short term.
[0050] In one example, the electronic device first obtains the identification information (such as IMEI number) of the user's current terminal through a data interface, and then obtains information such as terminal brand, model, launch time, and price range by matching with the TAC code terminal identification library. Subsequently, it pulls the user's corresponding multi-dimensional replacement data from contract management data, bundled package data, family association data, user profile data, and billing data. Finally, according to the preset mapping rules, the replacement data of each dimension is normalized and mapped to the corresponding replacement factor.
[0051] In some examples, multidimensional replacement factors include device age, average monthly user income, data traffic, terminal tier, age, integration and binding degree, contract lock-in period, and household ripple effect.
[0052] Among them, the integration binding factor reflects the depth of the SIM card binding between the current terminal application and the operator's integrated services; the contract lock-in period factor reflects the remaining time of the operator's contract bound to the current terminal; the family ripple factor reflects the strength of the mutual influence of the replacement behavior in the family corresponding to the current terminal; the device age factor reflects the age of the current terminal; the average monthly user revenue factor reflects the average monthly business revenue obtained by the operator from users; the data traffic factor reflects the data traffic usage corresponding to the current terminal; the terminal tier factor reflects the market value of the current terminal; and the age factor reflects the age of the user.
[0053] In some cases, multidimensional device replacement data includes: device age, average monthly revenue per user (ARPU), data usage, age, device tier, bundled information, contract lock-in period, and number of times family members replace devices within a preset time period.
[0054] In some cases, the equipment age is normalized and mapped to the range of 0-0.9 according to a preset mapping rule corresponding to the equipment age, resulting in an equipment age factor. For example, the preset mapping rule for equipment age is as follows: the equipment age is segmented according to months 0-12, 13-24, 25-36, 37 and above, and the equipment age of the corresponding segments is mapped to equipment age factors of 0.1, 0.4, 0.7 and 0.9 respectively.
[0055] In some examples, the average monthly user income is normalized and mapped to a range of 0-0.8 according to a preset mapping rule, resulting in an average monthly user income factor. For example, the preset mapping rule for average monthly user income is as follows: the average monthly user income is segmented into segments of less than 20 yuan, 20-50 yuan, 50-100 yuan, and 100 yuan and above, and the corresponding segments are mapped to average monthly user income factors of 0.3, 0.8, 0.7, and 0.5, respectively.
[0056] In some cases, traffic usage data is normalized and mapped to a range of 0-0.7 according to a preset mapping rule, resulting in a traffic factor. For example, the preset mapping rule for traffic usage data is as follows: traffic usage data is segmented into segments based on being less than a first traffic value, between the first and second traffic values, and greater than the second traffic value, and the traffic usage data in the corresponding segments is mapped to traffic factors of 0.7, 0.3, and 0.5 respectively.
[0057] In some cases, the age is normalized and mapped to the range of 0-0.9 according to the preset mapping rules corresponding to age, thus obtaining the age factor.
[0058] In some cases, the terminal gear is normalized and mapped to a range of 0.5-1.5 according to the preset mapping rules corresponding to the terminal gear, thus obtaining the terminal gear factor. For example, the terminal gears include entry-level, low-mid, standard, mid-high, and high-end. The preset mapping rules corresponding to the terminal gears are as follows: according to entry-level, low-mid, standard, mid-high, and high-end, the terminal gears are mapped to terminal gear factors of 1.5, 1.3, 1.0, 0.8, and 0.5 respectively.
[0059] In some cases, the fusion binding information is normalized and mapped to a value between 0.2 and 0.8 according to a preset mapping rule, resulting in a fusion binding factor. For example, the fusion binding information includes no fusion, single fusion, dual fusion, and triple fusion. The number of fusions refers to the number of broadband, TV, and mobile phones bound to the SIM card currently used by the user terminal. The preset mapping rule for the fusion binding information is as follows: according to no fusion, single fusion, dual fusion, and triple fusion, the fusion binding information is mapped to fusion binding factors of 0.2, 0.4, 0.5, and 0.8 respectively.
[0060] In some cases, the contract lock-up period is normalized and mapped to a range of 0.3-0.9 according to a preset mapping rule, resulting in a contract lock-up period factor. For example, the preset mapping rule for the contract lock-up period is as follows: the contract lock-up period is segmented into expired, within 6 months, 6-12 months, and 12 months and above, and the corresponding contract lock-up periods are mapped to contract lock-up period factors of 0.3, 0.5, 0.7, and 0.9 respectively.
[0061] In some examples, the number of times a family member switches devices within a preset time period is normalized and mapped to a range of 0.2-0.8 according to a preset mapping rule, thus obtaining the family ripple factor. For example, the preset mapping rule for the number of times a family member switches devices within a preset time period is as follows: the number of times a family member switches devices within a preset time period is segmented into 0 times, 1 time, 2 times or more, and the number of times a family member switches devices within the preset time period in the corresponding segments is mapped to family ripple factors of 0.2, 0.5, and 0.8 respectively.
[0062] S102, calculates the replacement index for users replacing their current terminals based on multi-dimensional replacement factors.
[0063] In some examples, multidimensional replacement factors include device age, average monthly user income, data usage, terminal tier, age, integration binding factor, contract lock-in period factor, and family ripple factor. Among these, the integration binding factor, contract lock-in period factor, and family ripple factor are all lock-in factors. Based on these multidimensional replacement factors, the replacement index for users to switch to their current terminal is calculated. This includes: weighting and summing the replacement factors (excluding the multidimensional lock-in factor) and subtracting the weighted sum of the multidimensional lock-in factor from the result to obtain a preliminary replacement index; calibrating and normalizing the preliminary replacement index using the calibration coefficient corresponding to the current terminal to obtain the final replacement index.
[0064] In some cases, the initial replacement index is calculated as follows:
[0065] in, , , , , , , , The factors are, in order: device age factor, average monthly user income factor, traffic factor, terminal tier factor, age factor, integration binding factor, contract lock-in period factor, and family ripple factor. , , , , , , , These are the preset weighting coefficients for the corresponding replacement factors.
[0066] In some cases, the preset weight coefficients of the device replacement factor are determined by extracting eight-dimensional samples of device replacement factors from historical device replacement databases for both users who have replaced their devices and those who have not. The samples are labeled with whether or not the device was replaced within a preset time period, and a logistic regression model is trained. The regression coefficients are then normalized and used as the weights for each dimension.
[0067] For example, , , , , , , , The values are 0.30, 0.15, 0.10, 0.10, 0.10, 0.10, 0.10, 0.05, respectively.
[0068] In some cases, the calibration coefficient is calculated as shown in Formula 1 below: (1) in, For calibration coefficients; This represents the percentage of users who have replaced their devices with the current generation of products within the current terminal's product series. This represents the percentage of users who have replaced their devices with the previous generation of the current terminal's series. , All are preset values greater than 0, and ; Indicates will The value is limited to .
[0069] In some examples, , The values are 0.5 and 2.0, respectively.
[0070] In some cases, after calibrating and normalizing the initial replacement index, the specific value of the replacement index is in the range of 0-100.
[0071] S103, calculate the time window for the user to change the current terminal based on the multidimensional locking factor. The size of the time window is negatively correlated with the value of any locking factor.
[0072] In some examples, the multidimensional locking factors include the fusion binding factor, the contract lock-in period factor, and the family ripple factor. The time window for a user to switch devices when changing their current terminal is calculated based on these multidimensional locking factors, including: calculating the time window according to the following formula 2-3: (2) (3) in, , , In order, they are: integration binding factor, contract lock-up period factor, and family ripple factor; , , In order , , The preset weight parameters; To lock in the efficiency score; The remaining time (the remaining time of the operator contract currently bound to the terminal). This is the preset replacement delay baseline.
[0073] In some cases, the replacement delay baseline Calibration: The median is taken from the sample distribution of historical contract expiration months to actual replacement months. The calibration value is used as the preset replacement delay baseline. For example, The value is 3 months.
[0074] S104 generates replacement prediction results based on the replacement index and replacement time window.
[0075] In some cases, the replacement index and replacement time window can be used directly to form the replacement prediction result.
[0076] In some examples, such as Figure 2 As shown, the SIM card currently used by the terminal is the primary card in the household; based on the replacement index and replacement time window, replacement prediction results are generated, including: S1041-S1044.
[0077] S1041, convert the replacement index into a group level according to the preset first mapping rule; In some examples, the preset first mapping rule can be to divide the replacement index into high-potential group, medium-potential group, and low-potential group according to the following categories: greater than the first value, between the first and second values, and less than the second value.
[0078] S1042, obtain the family binding degree corresponding to the family and the device replacement time window corresponding to each secondary card in the family.
[0079] In some examples, obtaining the family binding degree corresponding to a family may include: obtaining family binding information, which includes at least one of the following: the relationship between family members in the integrated package, the frequency of two-way calls, the co-occurrence of locations on holidays, and information on shared payment accounts; and mapping and normalizing the family binding information into a family binding degree according to a preset third mapping rule.
[0080] The calculation method for the device replacement time window corresponding to each secondary card can be found in Formula 2-3.
[0081] S1043, calculate the ripple propagation probability of a user switching devices to each family member corresponding to a secondary SIM card according to the following formula 4-6.
[0082] (4) (5) (6) in, For users, replacing the device is the first priority in the family. The ripple propagation probability corresponding to each secondary card; This is the ripple propagation amplification factor; For family binding degree; It is a cluster adjustment factor obtained by mapping the cluster level according to the preset second mapping rule; For the first The corresponding device replacement time window for each secondary SIM card. This represents the device replacement time window corresponding to the current terminal. This refers to the time window difference for equipment replacement. This is the attenuation coefficient.
[0083] In some cases, the ripple amplification factor The calibration method includes: from historical household phone replacement samples, statistically analyzing the frequency of secondary cards following the primary card replacement within a certain period (e.g., 6 months), and calculating the conditional probability separately for direct relatives and collateral relatives, which serves as the calibration value for β. For example, the β for direct relatives is 0.3, and the β for collateral relatives is 0.1.
[0084] In some cases, the attenuation coefficient The calibration methods include: selecting the model with the highest prediction accuracy from the historical distribution of household device replacement time differences. Value. For example, It lasts for 6 months.
[0085] In some examples, the second mapping rule sets the group adjustment factors for the high-potential group, medium-potential group, and low-potential group to be 1.2, 1.0, and 0.6, respectively.
[0086] S1044, generate a device replacement prediction result based on the ripple propagation probability, cluster level, and device replacement time window of the family members corresponding to each sub-card.
[0087] In some examples, multidimensional replacement factors also include device age factors, average monthly user revenue factors, data traffic factors, and terminal tier factors, such as... Figure 3 As shown, based on the ripple propagation probability, cluster level, and current terminal switching time window of each sub-card's family member, a switching prediction result is generated, which may include: S31-S32.
[0088] S31 determines the user's device replacement direction based on the device age factor, average monthly user income factor, traffic factor, and terminal tier factor. The replacement direction can be one of upgrading, remaining at the same tier, or downgrading.
[0089] In some cases, the first condition for upgrading is: and Upgrade condition two: and Downshifting conditions: or In all other cases, the price is flat.
[0090] S32 generates a device replacement prediction result based on the ripple propagation probability, cluster level, current terminal replacement time window, and replacement direction of the family members corresponding to each sub-card.
[0091] By utilizing multidimensional replacement factors that represent replacement intentions, including multidimensional locking factors that represent replacement resistance, the final replacement prediction result is obtained. This approach characterizes user replacement behavior from two main perspectives: whether users want to replace their devices and whether they are able to replace them. Compared to using a fixed threshold based on device age to predict replacements, this method reduces the false positive rate and increases the accuracy rate of prediction results.
[0092] Furthermore, by using multidimensional locking factors to calculate the user's device replacement time window and limiting the window size of the replacement time window to be negatively correlated with the value of any locking factor, it may seem to violate the rule that the greater the user's resistance to device replacement, the larger the replacement time window. However, this is actually an accurate model of user device replacement behavior based on a large amount of data analysis, summarization, and discovery. It can more accurately represent the relationship between the user's device replacement window and the multidimensional locking factors, thereby more accurately predicting the user's device replacement time window. Furthermore, limiting the data source to operator-side data ensures comprehensive coverage, high accuracy, and compliance with regulations, thus avoiding sensitivity issues when obtaining data from other sources.
[0093] Example 2: Another embodiment of this application relates to a terminal replacement prediction device. The implementation details of this embodiment's terminal replacement prediction device are described below. The following implementation details are provided for ease of understanding and are not essential for implementing this solution. A schematic diagram of this embodiment's terminal replacement prediction device can be seen as follows: Figure 4 As shown, it includes: a determination module 41, used to determine a multi-dimensional replacement factor based on the operator-side data corresponding to the user's current terminal, the multi-dimensional replacement factor including a multi-dimensional locking factor to reflect the user's resistance to replacing the terminal; a calculation module 42, used to calculate the replacement index of the user replacing the current terminal based on the multi-dimensional replacement factor; and to calculate the replacement time window of the user replacing the current terminal based on the multi-dimensional locking factor, the window size of the replacement time window being negatively correlated with the value of any locking factor; and a generation module 43, used to generate a replacement prediction result based on the replacement index and the replacement time window.
[0094] In some optional embodiments, the multidimensional locking factors include a convergence binding degree factor, a contract lock-up period factor, and a family ripple factor. The convergence binding degree factor is used to reflect the depth of the binding between the SIM card of the current terminal application and the operator's converged service. The contract lock-up period factor is used to reflect the remaining time of the operator contract bound to the current terminal. The family ripple factor is used to reflect the strength of the mutual influence of the switching behavior in the family corresponding to the current terminal. The calculation module 42 is used to calculate the switching time window according to formula 2-3.
[0095] In some optional embodiments, the SIM card of the current terminal application is the primary card in the family; the generation module 43 is used to convert the phone replacement index into a grouping level according to a preset first mapping rule; obtain the family binding degree corresponding to the family and the phone replacement time window corresponding to each secondary card in the family; calculate the ripple propagation probability of the user's phone replacement on the family members corresponding to each secondary card according to the following formulas 4-6; generate the phone replacement prediction result based on the ripple propagation probability of the family members corresponding to each secondary card, the grouping level and the phone replacement time window of the current terminal.
[0096] In some optional embodiments, the multidimensional replacement factors also include device age factor, average monthly user revenue factor, data traffic factor, and terminal tier factor. The device age factor reflects the current terminal's age, the average monthly user revenue factor reflects the operator's average monthly revenue from users, the data traffic factor reflects the current terminal's data traffic usage, and the terminal tier factor reflects the current terminal's market value. The generation module 43 determines the user's replacement direction based on the device age factor, average monthly user revenue factor, data traffic factor, and terminal tier factor. The replacement direction can be one of upgrading, remaining at the same tier, or downgrading. The module generates a replacement prediction result based on the ripple propagation probability, cluster level, current terminal replacement time window, and replacement direction of each secondary SIM card's corresponding family members.
[0097] In some optional embodiments, the multidimensional replacement factors also include device age factor, average monthly user income factor, traffic factor, terminal tier factor, and age factor. The multidimensional locking factors include integration binding factor, contract lock-up period factor, and family ripple factor. The age factor is used to reflect the user's age. The calculation module 42 is used to perform a weighted summation of the replacement factors other than the multidimensional locking factors in the multidimensional replacement factors, and subtract the weighted summation of the multidimensional locking factors from the obtained value to obtain a preliminary replacement index. The preliminary replacement index is calibrated and normalized using the calibration coefficient corresponding to the current terminal to obtain the replacement index.
[0098] By utilizing multidimensional replacement factors that represent replacement intentions, including multidimensional locking factors that represent replacement resistance, the final replacement prediction result is obtained. This approach characterizes user replacement behavior from two main perspectives: whether users want to replace their devices and whether they are able to replace them. Compared to using a fixed threshold based on device age to predict replacements, this method reduces the false positive rate and increases the accuracy rate of prediction results.
[0099] Furthermore, by using multidimensional locking factors to calculate the user's device replacement time window and limiting the window size of the replacement time window to be negatively correlated with the value of any locking factor, it may seem to violate the rule that the greater the user's resistance to device replacement, the larger the replacement time window. However, this is actually an accurate model of user device replacement behavior based on a large amount of data analysis, summarization, and discovery. It can more accurately represent the relationship between the user's device replacement window and the multidimensional locking factors, thereby more accurately predicting the user's device replacement time window. Furthermore, limiting the data source to operator-side data ensures comprehensive coverage, high accuracy, and compliance with regulations, thus avoiding sensitivity issues when obtaining data from other sources.
[0100] It is worth mentioning that all modules involved in this embodiment are logical modules. In practical applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. Furthermore, to highlight the innovative aspects of this application, this embodiment does not introduce units that are not closely related to solving the technical problems proposed in this application; however, this does not mean that other units are absent in this embodiment.
[0101] Example 3: Another embodiment of this application relates to an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the terminal switching prediction method in the above embodiments.
[0102] The memory and processor are connected via a bus, which can include any number of interconnecting buses and bridges, connecting various circuits of one or more processors and memories. The bus can also connect various other circuits, such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and will not be described further herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by the processor is transmitted over the wireless medium via an antenna, which further receives data and transmits it to the processor.
[0103] The processor manages the bus and general processing, and also provides various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory is used to store data used by the processor during operation.
[0104] Example 4: Another embodiment of this application relates to a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the method embodiments described above.
[0105] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0106] Example 5: Another embodiment of this application relates to a computer program product, including a computer program. When executed by a processor, the computer program implements the method embodiments described above.
[0107] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.
[0108] Furthermore, although the steps of the method in this application are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.
[0109] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware.
[0110] Those skilled in the art will understand that the above embodiments are specific embodiments for implementing this application, and in practical applications, various changes can be made to them in form and detail without departing from the spirit and scope of this application.
Claims
1. A method for predicting terminal replacement, characterized in that, include: A multi-dimensional device replacement factor is determined based on the operator-side data corresponding to the user's current terminal. The multi-dimensional device replacement factor includes a multi-dimensional locking factor used to reflect the user's resistance to device replacement. Based on the multidimensional replacement factor, calculate the replacement index for the user to replace the current terminal; The switching time window for a user to change the current terminal is calculated based on the multidimensional locking factor, and the window size of the switching time window is negatively correlated with the value of any locking factor. Based on the replacement index and the replacement time window, a replacement prediction result is generated.
2. The method according to claim 1, characterized in that, The multidimensional locking factors include a convergence binding factor, a contract lock-up period factor, and a family ripple factor. The convergence binding factor reflects the depth of the binding between the SIM card of the current terminal application and the operator's converged services. The contract lock-up period factor reflects the remaining time of the operator's contract bound to the current terminal. The family ripple factor reflects the strength of the mutual influence of handset switching behaviors in the family corresponding to the current terminal. The step of calculating the user's device replacement time window based on the multidimensional locking factor includes: The replacement time window is calculated using the following formula: in, , , The factors are, in order: fusion binding factor, contract lock-in period factor, and family ripple factor; , , In order , , The preset weight parameters; To lock in the efficiency score; The remaining time; This is the preset replacement delay baseline.
3. The method according to claim 1, characterized in that, The SIM card used in the current terminal application is the primary card in the household; the generation of a phone replacement prediction result based on the phone replacement index and the phone replacement time window includes: The replacement index is converted into a cluster level according to a preset first mapping rule; Obtain the family binding degree corresponding to the family and the device switching time window corresponding to each secondary SIM card in the family; The ripple propagation probability of the user switching devices to each family member corresponding to the secondary SIM card is calculated using the following formula: in, For the user to replace the device, the first in the household The ripple propagation probability corresponding to each secondary card; This is the ripple propagation amplification factor; The family binding degree; It is a cluster adjustment factor obtained by mapping the cluster level according to a preset second mapping rule; For the first The corresponding device replacement time window for each secondary SIM card. This refers to the device switching time window corresponding to the current terminal. This refers to the time window difference for equipment replacement. The attenuation coefficient; The device switching prediction result is generated based on the ripple propagation probability of family members corresponding to each sub-card, the grouping level, and the device switching time window of the current terminal.
4. The method according to claim 3, characterized in that, The multi-dimensional replacement factor also includes device age factor, average monthly user revenue factor, data traffic factor, and terminal tier factor. The device age factor reflects the age of the current terminal, the average monthly user revenue factor reflects the average monthly business revenue that the operator receives from the user, the data traffic factor reflects the data traffic usage of the current terminal, and the terminal tier factor reflects the market value of the current terminal. The step of generating the device switching prediction result based on the ripple propagation probability of family members corresponding to each secondary card, the grouping level, and the switching time window of the current terminal includes: Based on the device age factor, the average monthly user income factor, the traffic factor, and the terminal tier factor, the user's device replacement direction is determined, which is one of upgrading, remaining at the same tier, or downgrading. The device switching prediction result is generated based on the ripple propagation probability of the family members corresponding to each sub-card, the grouping level, the device switching time window of the current terminal, and the device switching direction.
5. The method according to claim 1, characterized in that, The multi-dimensional replacement factors also include device age factor, average monthly user income factor, traffic factor, terminal tier factor, and age factor. The multi-dimensional locking factors include integration binding factor, contract lock-in period factor, and family ripple factor. The age factor is used to reflect the user's age. The step of calculating the replacement index for the user to replace the current terminal based on the multidimensional replacement factor includes: The weighted summation of the replacement factors other than the multidimensional locking factor in the multidimensional replacement factors is performed, and the weighted summation of the multidimensional locking factor is subtracted from the obtained value to obtain the preliminary replacement index. The initial replacement index is calibrated and normalized using the calibration coefficient corresponding to the current terminal to obtain the replacement index.
6. The method according to claim 5, characterized in that, The calibration coefficient is calculated as follows: in, The calibration coefficient is mentioned above; This represents the cumulative replacement rate of current-generation products within the current terminal's product series. The percentage of users who have replaced their devices with the previous generation of the series to which the current terminal belongs; , All are preset values greater than 0, and ; Indicates will The value is limited to .
7. A terminal replacement prediction device, characterized in that, include: The determination module is used to determine a multi-dimensional device replacement factor based on the operator-side data corresponding to the user's current terminal. The multi-dimensional device replacement factor includes a multi-dimensional locking factor that reflects the user's resistance to device replacement. The calculation module is used to calculate the replacement index of the user replacing the current terminal based on the multidimensional replacement factor; The switching time window for a user to change the current terminal is calculated based on the multidimensional locking factor, and the window size of the switching time window is negatively correlated with the value of any locking factor. The generation module is used to generate replacement prediction results based on the replacement index and the replacement time window.
8. An electronic device, characterized in that, include: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the terminal replacement prediction method as described in any one of claims 1 to 6.
9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the terminal replacement prediction method according to any one of claims 1 to 6.
10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the terminal replacement prediction method according to any one of claims 1 to 6.
